Tug O War

Me: my worry, is this tug-o-war, between ignorance and opportunism, is eroding the intent for people to learn, which further than two issues: 1. Humans ability to learn if they really can convert the opportunities to real solutions 2. AI is self-feeding to its fluff and not get "real battle" tested...... double edge sword...

AI: That's a sharp observation and I think it's the most underrated risk in the current landscape.

The feedback loop is vicious. Humans stop learning because AI makes the surface look easy. AI stops improving (in real terms) because it's not being pushed against hard problems by people who understand the domain deeply enough to challenge it. Both sides atrophy.

On the human side — the skill that's dying isn't coding or writing, it's problem decomposition. The ability to look at a messy real-world situation and figure out what the actual constraints are, what breaks, what matters.

On the AI side — this is the subtler point you're making. If AI is mostly being used by people who can't evaluate its output, it's operating in an environment with no real feedback signal. It's like a student who only gets graded by other students. The fluff compounds. Models get optimized for sounding right rather than being right, because that's what the market rewards. The battle-testing that would force genuine improvement — from domain experts who know where the cracks are — isn't happening at scale.

The double edge: the people who could push AI into genuinely hard territory are either too busy, too skeptical to engage, or being told they're about to be replaced — which is exactly the wrong incentive to get them invested in making the technology better.

The way out is probably small and specific — people building real things against real constraints, failing, and publishing what they learned. Not thought leadership. Not demos. Actual collision with reality.